Saturday, May 27, 2006

Paradise Lost

Back to infrequent posts here - the commute to London for the BT 'Wireless Cities' project takes up a great deal of time.

Down to the Salisbury Playhouse this afternoon to see 'Paradise Lost' (John Milton) adapted for stage by Ben Power. Press release here.

The first half has Satan and cohorts planning their future strategy, having ended up in hell after losing a power struggle with God. Cue lightshow and loud music.

After the interval, we are in the Garden of Eden with a naked Adam and Eve on stage, about to be tempted by the serpent-disguised Satan. The play ends with A. and E.'s expulsion from the G. of E. following the apple incident.

Although 'highly-aclaimed', we were not unduly impressed. The language is sub-Shakespearean, and the acting - although technically competent - seems to lack conviction. The central problem, however, is with the plot. The dilemma facing A. and E. is to eat the apple, get knowledge and take charge of their destiny (but incur God's wrath for disobedience), vs. obey God's arbitrary injunction not to eat it and continue to wallow in brain-dead luxury.

Since there is no rational reason for God to have forbidden the apple to be eaten, or at least no motive which the script identifies, there is no interesting central dilemma for the characters. And hence no dramatic tension.

Confession: we both almost fell asleep in the first half, and we did not find the nudity arousing in the second. Looks like the naturists are right after all.

Monday, May 22, 2006

BT Wireless City project

Attended a meeting in London earlier today where I signed up for BT's wireless city project. This will roll-out municipal WiFi in a number of UK cities, in a pattern previously seen in San Francisco and Philadelphia. I will be managing roll-out in a couple of the cities (yet tbd) through my company Interweave Consulting.

Just as well that I shipped the book manuscript to the publisher on Friday.

Monday, May 15, 2006

Really finished this time!

Over the week-end I really blitzed the last chapter (12) on 'spoken dialogue systems' and produced 5,000 words, which is OK. This is now out for review - it's titled 'Machines Who Talk'.

My other major task was to incorporate the anonymous reviewer's comments, which mostly involved adding material about GMPLS, Grid Computing, SONET and a review of current 'Internet2'-type systems.

I am now just fine-tuning and rewriting the odd point here and there, and tomorrow I will review again the last three chapters. Once comments are back from the various people who are looking at material, I think this is about ready to ship :-)

Wednesday, May 10, 2006

Data Bombing

Something weird affecting my laptop.

It has an 18.6 GB hard disk, with 5.3 GB free. However, Monday evening (May 8th) I got a message that the disk was full. Yep, only a few Megabytes left. When I checked folders very carefully, I found that the WINDOWS/TEMP folder was filled with files like this:

AcrF4.tmp, AcrF5.tmp, ..., AcrFF.tmp, Acr100.tmp, Acr101.tmp, ... etc. Counting up in hex.

Each file was around 4 MB in size (just under, just over, it depended) and they were being written into the folder around four times a minute, although the rate was also variable. Windows marked the application as unknown, and when I opened a file in Notepad, it was binary.

Aty first I suspected BOINC, the distributed science application platform. But I stopped it, and the problem continued. I checked the active processes, and nothing seemed amiss. It's a shame there is no disk access logger with XP though.

In desperation I have re-installed the McAfee antivirus program (I had some suspicions) and have used System Restore to back up to last Sunday. As of this time, the Acr files have stopped arriving. More later (if the problem recurs).

UPDATE: Thursday morning, May 11th 06

End of last evening, the problem had not gone away. This morning, I started by closing down systematically all my user processes (via the task manager) - checking at each stage whether the demonic file-writing would stop. It did not.

I then downloaded the freeware program: mst IsUsedBy, a utility which opens a window, into which you can drag a file. The program then tells you which process is using that file. It told me that the culprit was the Microsoft indexing program cidaemon.exe. Surprising, because although this program can apparently be a CPU hog, it is not flagged as dumping data.

I tracked it down to C:WINDOWS/system32 and renamed the process (prepended an 'x'). This may have cured the problem. However, re-checking in this folder, I see that Windows has put it back! Still ...

I also took the precaution of organising all my data, collecting my executables together, listing the programs I continue to use, and backing everything up to my external HD. This, because the next step is a full Windows XP re-install on the assumption that I have a piece of malware here. It has evaded McAfee antivirus, spybot S&D and Lavasoft ad-aware so it's pretty pernicious.

A re-install from Toshiba's 2003 CDs looks like an all-day operation (SP2 download all over again, etc) so it really is a last resort...

UPDATE Friday May 12th 9.00 a.m.

Now 24 hours without any reappearance of the problem. I guess we can say fixed. What seems to have happened is that the cidaemon.exe process appears to have gone 'rogue'. Whether this was a bit error, or some corruption of a config file or debug setting is completely unknown.

However, Windows recreated the process, presumably from a clean source, and the problem has gone away. By side effect, it did force a clean-up of my data and an audit of what I have as well as rehearsing the complete recovery process - which has value. Still, what a waste of time, overall.

Tuesday, May 09, 2006

New Toy

The Sony-Ericsson K750i camera phone I bought a few days ago (see previous blog) is proving pretty addictive. It's just too easy to take quick home movies, although the resolution is very poor. The subsequent process is slightly messy.

1. The movie is transferred to the PC, emulating a USB drive - in .3gp format.

2. I then map it to .avi using the tool I bought (previous post).

3. I then import this into Windows Movie Maker, which allows the weird multi-clip format from the camera phone to be smoothed out into one movie, with title if required. This is saved in .wmv format (no choice! That's why Microsoft develped the tool for free).

And then the home movie is inflicted on anyone I misguidedly believe might be interested. Strange that you can know this, and still do it.

By putting the movies on my website (a minute comes in just under 5 MB) at least I send a link rather than the file. And Windows Media Player buffers and plays on download, rather than waiting for the entire file to be downloaded, so it's almost like streaming ...

No, I can't resist ... our visit last week-end to Wardour castle, destroyed in the English civil war - here at 1.8 MB (30 seconds).

Friday, May 05, 2006

The NGN in 500 words

When I meet with clients, I am sometimes asked to explain briefly what this NGN thing is all about. Sadly, they don’t get a snappy answer because there isn’t one. Here is my attempt to provide the simplest, most concise answer which is accurate enough to be useful.

The Next-Generation Network (NGN) is first of all a technical innovation in three parts.

1. A new version of the Internet with added gold, silver, bronze service classes. (IP/MPLS transport with Diffserv).

2. A multimedia phone service which can also be used to download music, video-clips and TV programming as well as make voice or video calls and check your house surveillance cameras. (IMS).

3. A carrier platform which can be used by companies to supply application services over the Internet. For example, where companies now have to install a lot of software to run their businesses, in future their employees can just point their browsers at Internet sites which will supply these business functions as a service. These sites will be provided by ASPs - Application Service Providers (not quite the same as e-Retailers like Amazon, which sell books and things). Even desktop functions like word processing, slide-pack creation and spreadsheets will be available as Internet web services, as an alternative to running programs on your PC. (Hosted Java EE or .NET platforms + AJAX)

The NGN itself is really middleware. To make it profitable, it has to be used to provide new services. But powerful companies like Microsoft, Google, Yahoo, Sky and the BBC already have applications and content which they will want to distribute over the NGN. Will they use the carriers to do this, possibly via an alliance, or will they forwards-integrate into the carrier business themselves?

The carriers have a similar dilemma. Should they become ASPs, Internet Portals and TV broadcasters, should they try to extract value (rents) from the Googles, Yahoos and Skys, or will attempts to do so simply persuade these powerful upstream players to sidestep the carriers and build their own NGNs?

NGN services also have to be delivered to customers. Simple services like those delivered on mobile phones today require just a straightforward retail operation. But delivering complex business solutions to enterprises requires business and systems analysis and design, and complex implementation and configuration. Value-Added Resellers (VARs) and Systems Integrators (SIs) do these jobs today, often in competition with the carriers. Should the carriers beef-up their professional services arms? If not, do they risk commoditisation?

And finally, the investment in NGNs is really in three parts.

1. Buy new network and IT equipment.

2. Redesign the organisation’s products, processes and roles, and migrate from the old.

3. Close down the legacy company and release large numbers of redundant staff.

The reduction in headcount is where the real saving come from. However, to accomplish steps 1-3, especially 2, requires a huge expenditure of both capital and effort. It is probably beyond the abilities of ‘old alternate operators’, those who still have legacy networks today. Are they doomed? Probably, unless someone with very deep pockets can find a reason to buy and fix them.

The world after the NGN has all the big players: ASPs, content-aggregators like Sky, sleek and large NGN carriers, SIs and VARs. Things will be open enough to allow smaller, more innovative plays, but competition will be tougher. After 2010, telecoms will look more like IT today.

To probe further, check here.

Wednesday, May 03, 2006

Book Review: Computational Linguistics

Posted at here.

The Oxford Handbook of Computational Linguistics (Oxford Handbooks) (Paperback).

This ‘handbook’ needs both hands to lift it! At 700+ pages and 38 chapters, detailed chapter-by-chapter review is impossible. Let me start with the top-level structure, which divides the book into three parts: Fundamentals; Processes, Methods and Resources; and Applications.

Part one, ‘Fundamentals’, walks through the standard sub-disciplines of computational linguistics with chapter headings: phonology, morphology, lexicography, syntax, semantics, discourse, pragmatics and dialogue, formal grammars and languages, complexity theory. Each chapter is a short introduction and overview to the topic, aimed at the informed newcomer (i.e. it helps if you have a computer science/maths background and know about predicate logic and state machines).

Part two, ‘Processes, etc’, covers a number of problem areas and techniques: text-segmentation, part-of-speech tagging, parsing, word-sense disambiguation, anaphora resolution, natural language generation and so on. There is little commonality between the chapters, but they are all informative.

The final part, ‘Applications’ covers areas such as machine translation, information retrieval, text summarisation, second-language computer-assisted learning systems and spoken dialogue systems.

As a comprehensive, and relatively recent review of the whole field the book is excellent. Some points which caught my interest.

1. Speech and written language are hugely different, due to noise, self-repair, speech acts and discourse functions, accents and the strange ‘grammaticality’ of utterances (p. 521).

2. The distinction between simpler finite-state dialogue models (machine-centric) vs. more dynamic planning-based dialogue managers (which can deal with mixed-initiative dialogue) - chapter 7.

3. The controversial role of real-world knowledge. This is different from semantics, which is more about representational and inferential adequacy. Chapter 25 on Ontologies surprising states “it is not clear to what extent NLP technology, in its current form, needs such ontologies and their complex knowledge representation systems”. Apparently “large scale vocabularies with very limited reasoning are preferred”. Interesting.

Human-to-human conversation seems, in performance, to be a unitary phenomenon. For scientific purposes, however, it has to be analysed into sub-fields, as in the chapter headings of part one. However, there is then both the problem of tunnel vision, and of scope creep: we see, for example, syntactic approaches expanding into the spaces of semantics and pragmatics in, to my mind, an unbalanced way.

I was most interested in Spoken Dialogue Systems, as these attempt to combine the state of the art in the separate disciplines into a unified architecture and implementation to address the original problem: a powerful constraint on one-sided development. The solution architectures seem to show that modular works, with bottom-up statistical techniques performing well at the speech-recognition level, and symbolic processing techniques such as automatic planning to achieve agent goals working at the dialogue level. The latter seems to be the least developed, however, as linguistics merges into a more general social agent theory.

Monday, May 01, 2006

Sony Ericsson K750i camera phone

I upgraded to this phone on Friday, and a very cute piece of equipment it is. Takes about four hours of hard work to get the hang of it, and then you have:

  • A very intuitive and high-quality FM radio
  • A 2 Megapixel still camera
  • A video-camera which also records sound
  • An MP3 player
  • A sound recorder
  • And of course it handles calls and SMS.

Comments. There is no view-finder, so you take pictures by looking at the screen. This can be hard on a sunny day, and worse if you are wearing shades. The quality, though, is excellent. The video camera takes unbounded-length video clips in 3gp format (I spent an extra £20 to buy the AVS video convertor tool set to map 3gp into more familar formats such as .avi and .mpeg). However, the 3gp format is extremely compressed - 30 seconds video in under 500 kB. The downside is that the resulting size is dimensioned to a mobile phone screen - just a little smaller than a credit card.

The USB cable linking to the PC makes the phone memory card look like a USB drive, (I bought a 512 MB card - currently available up to 2 GB), so moving MP3, jpeg and 3gp files around is just drag 'n' drop. All-in-all, a nice piece of equipment and when the 8GB memory card becomes available it will almost obsolete MP3 players (if they can figure a way to play music and have the phone off - needed in places like hospitals and aircraft).

We drove down to Chilbolton, five miles south-east of Andover, lunchtime today to take a look at the world's largest fully steerable meteorological radar, the Chilbolton Advanced Meteorological Radar (CAMRa) - site here. Unfortunately you can't visit, but the K750i provided the photograph above, cropped at the original resolution.